well-defined point or points of reference in the patterns. This was done by utilising the centres of the one or two triangular plots (see Plate 4, Fig. 8, ~2~, ~3~, ~4~) which are found in the great majority of patterns, and whose existence was pointed out by Purkenje, but not their more remote cause, which is as follows: The ridges, as was shown in the diagram (Plate 3) of the palm of the hand, run athwart the fingers in rudely parallel lines up to the last joint, and if it were not for the finger-nail, would apparently continue parallel up to the extreme finger-tip. But the presence of the nail disturbs their parallelism and squeezes them downwards on both sides of the finger. (See Fig. 8, ~2~.) Consequently, the ridges that run close to the tip are greatly arched, those that successively follow are gradually less arched until, in some cases, all signs of the arch disappear at about the level of the first joint (Fig. 8, ~1~). Usually, however, this gradual transition from an arch to a straight line fails to be carried out, causing a break in the orderly sequence, and a consequent interspace (Fig. 8, ~2~). The topmost boundary of the interspace is formed by the lowermost arch, and its lowermost boundary by the topmost straight ridge. But an equally large number of ducts exist within the interspace, as are to be found in adjacent areas of equal size, whose mouths require to be supported and connected. This is effected by the interpolation of an independent system of ridges arranged in loops (Fig. 8, ~3~; also Plate 5, Fig. 9, _a_, _f_), or in scrolls (Fig. 8, ~4~; also Fig. 9, _g_, _h_), and this interpolated system forms the "pattern." Now the existence of an interspace implies the divergence of two previously adjacent ridges (Fig. 8, ~2~), in order to embrace it. Just in front of the place where the divergence begins, and before the sweep of the pattern is reached, there are usually one or more very short cross-ridges. Their effect is to complete the enclosure of the minute triangular plot in question. Where there is a plot on both sides of the finger, the line that connects them (Fig. 8, ~4~) serves as a base line whereby the pattern may be oriented, and the position of any point roughly charted. Where there is a plot on only one side of the finger (Fig. 8, ~3~), the pattern has almost necessarily an axis, which serves for orientation, and the pattern can still be charted, though on a different principle, by dropping a perpendicular from the plot on to the axis, in the way there shown.